The hypoosmotic lysis curve of freshly collected human erythrocytes is
consistent with a single Gaussian error function with a mean of 46.5
+/- 0.25 mM NaCl and a standard deviation of 5.0 +/- 0.4 mM NaCl. Afte
r extended storage of RBCs under standard blood bank conditions the ly
sis curve conforms to the sum of two error functions instead of a poss
ible shift in the mean and a broadening of a single error function. Th
us, two distinct sub-populations with different fragilities are presen
t instead of a single, broadly distributed population. One population
is identical to the freshly collected erythrocytes, whereas the other
population consists of osmotically fragile cells. The rate of generati
on of the new, osmotically fragile, population of cells was used to pr
obe the hypothesis that lipid peroxidation is responsible for the indu
ction of membrane fragility. If it is so, then the antioxidant, tirila
zad mesylate (U-74,006f), should protect against this degradation of s
tored erythrocytes. We found that tirilazad mesylate, at 17 mu M (1.5
mol% with respect to membrane lecithin), retards significantly the for
mation of the osmotically fragile RBCs. Concomitantly, the concentrati
on of free hemoglobin which accumulates during storage is markedly red
uced by the drug. Since the presence of the drug also decreases the am
ount of F-2-isoprostxanes formed during the storage period, an antioxi
dant mechanism must be operative. These results demonstrate that tiril
azad mesylate significantly decreases the number of fragile erythrocyt
es formed during storage in the blood bank.